Medios-Macina/cmdnat/worker.py

"""Worker cmdlet: Display workers table in ResultTable format."""
from __future__ import annotations

import json
import sys
from dataclasses import dataclass
from datetime import datetime, timezone
from typing import Any, Dict, Sequence, List

from cmdlet import register
from cmdlet._shared import Cmdlet, CmdletArg
import pipeline as ctx
from SYS.logger import log
from config import get_local_storage_path

DEFAULT_LIMIT = 100
WORKER_STATUS_FILTERS = {"running", "completed", "error", "cancelled"}
HELP_FLAGS = {"-?", "/?", "--help", "-h", "help", "--cmdlet"}

CMDLET = Cmdlet(
    name=".worker",
    summary="Display workers table in result table format.",
    usage=".worker [status] [-limit N] [@N]",
    arg=[
        CmdletArg("status", description="Filter by status: running, completed, error (default: all)"),
        CmdletArg("limit", type="integer", description="Limit results (default: 100)"),
        CmdletArg("@N", description="Select worker by index (1-based) and display full logs"),
        CmdletArg("-id", description="Show full logs for a specific worker"),
        CmdletArg("-clear", type="flag", description="Remove completed workers from the database"),
    ],
    detail=[
        "- Shows all background worker tasks and their output",
        "- Can filter by status: running, completed, error",
        "- Search result stdout is captured from each worker",
        "- Use @N to select a specific worker by index and display its full logs",
        "Examples:",
        ".worker                        # Show all workers",
        ".worker running                # Show running workers only",
        ".worker completed -limit 50    # Show 50 most recent completed workers",
        ".worker @3                     # Show full logs for the 3rd worker",
        ".worker running @2             # Show full logs for the 2nd running worker",
    ],
)


def _has_help_flag(args_list: Sequence[str]) -> bool:
    return any(str(arg).lower() in HELP_FLAGS for arg in args_list)


@dataclass
class WorkerCommandOptions:
    status: str | None = None
    limit: int = DEFAULT_LIMIT
    worker_id: str | None = None
    clear: bool = False


@register([".worker", "worker", "workers"])
def _run(result: Any, args: Sequence[str], config: Dict[str, Any]) -> int:
    """Display workers table or show detailed logs for a specific worker."""
    args_list = [str(arg) for arg in (args or [])]
    selection_indices = ctx.get_last_selection()
    selection_requested = bool(selection_indices) and isinstance(result, list) and len(result) > 0

    if _has_help_flag(args_list):
        log(json.dumps(CMDLET, ensure_ascii=False, indent=2))
        return 0

    options = _parse_worker_args(args_list)

    library_root = get_local_storage_path(config or {})
    if not library_root:
        log("No library root configured", file=sys.stderr)
        return 1

    try:
        from API.folder import API_folder_store

        with API_folder_store(library_root) as db:
            if options.clear:
                count = db.clear_finished_workers()
                log(f"Cleared {count} finished workers.")
                return 0

            if options.worker_id:
                worker = db.get_worker(options.worker_id)
                if worker:
                    events: List[Dict[str, Any]] = []
                    try:
                        wid = worker.get("worker_id")
                        if wid and hasattr(db, "get_worker_events"):
                            events = db.get_worker_events(wid)
                    except Exception:
                        pass
                    _emit_worker_detail(worker, events)
                    return 0
                log(f"Worker not found: {options.worker_id}", file=sys.stderr)
                return 1

            if selection_requested:
                return _render_worker_selection(db, result)

            return _render_worker_list(db, options.status, options.limit)
    except Exception as exc:
        log(f"Workers query failed: {exc}", file=sys.stderr)
        import traceback

        traceback.print_exc(file=sys.stderr)
        return 1


def _parse_worker_args(args_list: Sequence[str]) -> WorkerCommandOptions:
    options = WorkerCommandOptions()
    i = 0
    while i < len(args_list):
        arg = args_list[i]
        low = arg.lower()
        if low in {"-limit", "--limit"} and i + 1 < len(args_list):
            options.limit = _normalize_limit(args_list[i + 1])
            i += 2
        elif low in {"-id", "--id"} and i + 1 < len(args_list):
            options.worker_id = args_list[i + 1]
            i += 2
        elif low in {"-clear", "--clear"}:
            options.clear = True
            i += 1
        elif low in {"-status", "--status"} and i + 1 < len(args_list):
            options.status = args_list[i + 1].lower()
            i += 2
        elif low in WORKER_STATUS_FILTERS:
            options.status = low
            i += 1
        elif not arg.startswith("-"):
            options.status = low
            i += 1
        else:
            i += 1
    return options


def _normalize_limit(value: Any) -> int:
    try:
        return max(1, int(value))
    except (TypeError, ValueError):
        return DEFAULT_LIMIT


def _render_worker_list(db, status_filter: str | None, limit: int) -> int:
    workers = db.get_all_workers(limit=limit)
    if status_filter:
        workers = [w for w in workers if str(w.get("status", "")).lower() == status_filter]

    if not workers:
        log("No workers found", file=sys.stderr)
        return 0

    for worker in workers:
        started = worker.get("started_at", "")
        ended = worker.get("completed_at", worker.get("last_updated", ""))

        date_str = _extract_date(started)
        start_time = _format_event_timestamp(started)
        end_time = _format_event_timestamp(ended)

        item = {
            "columns": [
                ("Status", worker.get("status", "")),
                ("Pipe", _summarize_pipe(worker.get("pipe"))),
                ("Date", date_str),
                ("Start Time", start_time),
                ("End Time", end_time),
            ],
            "__worker_metadata": worker,
            "_selection_args": ["-id", worker.get("worker_id")],
        }
        ctx.emit(item)
    return 0


def _render_worker_selection(db, selected_items: Any) -> int:
    if not isinstance(selected_items, list):
        log("Selection payload missing", file=sys.stderr)
        return 1

    emitted = False
    for item in selected_items:
        worker = _resolve_worker_record(db, item)
        if not worker:
            continue
        events: List[Dict[str, Any]] = []
        try:
            events = db.get_worker_events(worker.get("worker_id")) if hasattr(db, "get_worker_events") else []
        except Exception:
            events = []
        _emit_worker_detail(worker, events)
        emitted = True
    if not emitted:
        log("Selected rows no longer exist", file=sys.stderr)
        return 1
    return 0


def _resolve_worker_record(db, payload: Any) -> Dict[str, Any] | None:
    if not isinstance(payload, dict):
        return None
    worker_data = payload.get("__worker_metadata")
    worker_id = None
    if isinstance(worker_data, dict):
        worker_id = worker_data.get("worker_id")
    else:
        worker_id = payload.get("worker_id")
        worker_data = None
    if worker_id:
        fresh = db.get_worker(worker_id)
        if fresh:
            return fresh
    return worker_data if isinstance(worker_data, dict) else None


def _emit_worker_detail(worker: Dict[str, Any], events: List[Dict[str, Any]]) -> None:
    stdout_content = worker.get("stdout", "") or ""

    lines = stdout_content.splitlines()

    for line in lines:
        line = line.strip()
        if not line:
            continue

        timestamp = ""
        level = "INFO"
        message = line

        try:
            parts = line.split(" - ", 3)
            if len(parts) >= 4:
                ts_str, _, lvl, msg = parts
                timestamp = _format_event_timestamp(ts_str)
                level = lvl
                message = msg
            elif len(parts) == 3:
                ts_str, lvl, msg = parts
                timestamp = _format_event_timestamp(ts_str)
                level = lvl
                message = msg
        except Exception:
            pass

        item = {
            "columns": [
                ("Time", timestamp),
                ("Level", level),
                ("Message", message),
            ]
        }
        ctx.emit(item)

    # Events are already always derived from stdout for now.


def _summarize_pipe(pipe_value: Any, limit: int = 60) -> str:
    text = str(pipe_value or "").strip()
    if not text:
        return "(none)"
    return text if len(text) <= limit else text[: limit - 3] + "..."


def _format_event_timestamp(raw_timestamp: Any) -> str:
    dt = _parse_to_local(raw_timestamp)
    if dt:
        return dt.strftime("%H:%M:%S")

    if not raw_timestamp:
        return "--:--:--"
    text = str(raw_timestamp)
    if "T" in text:
        time_part = text.split("T", 1)[1]
    elif " " in text:
        time_part = text.split(" ", 1)[1]
    else:
        time_part = text
    return time_part[:8] if len(time_part) >= 8 else time_part


def _parse_to_local(timestamp_str: Any) -> datetime | None:
    if not timestamp_str:
        return None
    text = str(timestamp_str).strip()
    if not text:
        return None

    try:
        if "T" in text:
            return datetime.fromisoformat(text)
        if " " in text:
            dt = datetime.strptime(text, "%Y-%m-%d %H:%M:%S")
            dt = dt.replace(tzinfo=timezone.utc)
            return dt.astimezone()
    except Exception:
        pass
    return None


def _extract_date(raw_timestamp: Any) -> str:
    dt = _parse_to_local(raw_timestamp)
    if dt:
        return dt.strftime("%m-%d-%y")

    if not raw_timestamp:
        return ""
    text = str(raw_timestamp)
    date_part = ""
    if "T" in text:
        date_part = text.split("T", 1)[0]
    elif " " in text:
        date_part = text.split(" ", 1)[0]
    else:
        date_part = text

    try:
        parts = date_part.split("-")
        if len(parts) == 3:
            year, month, day = parts
            return f"{month}-{day}-{year[2:]}"
    except Exception:
        pass
    return date_part